157 Boundary Operation

A mapping that maps a domain to its edge that is used to relate a definite integral over a domain to a definite integral on that edge.

Note: Also written \(\partial\). Also called boundary operator.

definition [d] (Boundary Operation = Boundary Operator = \(\partial\)) From Emam: in more advanced discussions, the symbol \(\partial\) is treated as an operator—a sort of derivative—and it has the property of being nilpotent, which means that applying it twice always vanishes:

  • \(\partial^{2} = 0\) .

where

  • \(\partial\) is the boundary operation.
  • \(\partial M\) is the edge of a manifold or domain \(M\).
  • \(\partial^{2} = 0\) means the edge of an edge is empty.

definition [d] (Boundary Operation) From Gowers: in Stokes’s theorem,

  • \(\displaystyle \int_{S} d\omega = \int_{\partial S} \omega\)

for an oriented manifold \(S\) and form \(\omega\), where \(\partial S\) is the oriented boundary of \(S\). Differentiation \(\omega \mapsto d\omega\) is the adjoint of the boundary operation.

where

  • \(\partial S\) is the edge of \(S\) under the boundary operation.
  • \(d\omega\) is the exterior derivative of \(\omega\).

A mapping associates a function with another function. Under this mapping, the definite integral of a derivative over a domain has the same value as the definite integral of the function over the edge of the domain that is used to find a solution to an equation.

definition [d] (Boundary Operation) From Gowers, stated with elementary terms: an equation relates the definite integral of a derivative over a domain to the definite integral over its edge,

  • \(\displaystyle \int_{D} Df = \int_{\partial D} f\) .

where

  • both integrals are definite integrals.
  • \(D\) is the domain, the set on which the definite integral of the derivative is taken.
  • \(\partial D\) is the edge of the domain under the boundary operation.
  • \(f\) is the function whose values appear in the definite integral on the edge.
  • \(Df\) is the derivative of \(f\), which appears in the definite integral over the domain.
  • \(\partial\) denotes the boundary operation.

157.1 Examples

Before the boundary operation acts, one has a domain set \(A\). After the boundary operation acts, one has an edge set \(B = \partial A\). Stokes’s theorem then relates a definite integral over \(A\) to a definite integral over \(B\).

example 1 [d] (Closed interval to endpoints — Nash and Sen; Emam) Before: take the domain set

  • \(A = [a,b] = \{ x \in \mathbb{R} : a \le x \le b \}\) .

After: the boundary operation maps \(A\) to the edge set of endpoints

  • \(B = \partial A = \{ a, b \}\) .

Nash and Sen record the same edge for the half-open interval: if \(U = [a,b)\), then \(U^{\circ} = (a,b)\) and \(\overline{U} = [a,b]\), so \(b(U) = \overline{U} - U^{\circ} = \{ a, b \}\). Emam states the same passage of sets in manifold language: if \(M = [a,b]\), then \(\partial M\) has dimension zero and is the pair of points \(\{ a, b \}\).

where

  • \(A\) is the domain set before \(\partial\) acts.
  • \(B = \partial A\) is the edge set after \(\partial\) acts.
  • \(a\) and \(b\) are real numbers with \(a < b\).

example 2 [d] (Closed unit disk to circle — Lee) Before: take the domain set

  • \(A = \overline{B}^{2} = \{ x \in \mathbb{R}^{2} : |x| \le 1 \}\) .

After: the boundary operation maps \(A\) to the unit circle

  • \(B = \partial A = S^{1} = \{ x \in \mathbb{R}^{2} : |x| = 1 \}\) .

Lee: the closed unit disk \(\overline{B}^{2}\) is a manifold with boundary whose manifold boundary is the circle. Its topological boundary as a subset of \(\mathbb{R}^{2}\) is that same circle.

where

  • \(A\) is the domain set before \(\partial\) acts.
  • \(B = \partial A\) is the edge set after \(\partial\) acts.
  • \(|x|\) is the Euclidean length of \(x\).

example 3 [d] (Upper half-space to hyperplane — Lee) Before: take the domain set

  • \(A = \mathbb{H}^{n} = \{ (x_{1},\ldots,x_{n}) \in \mathbb{R}^{n} : x_{n} \ge 0 \}\) .

After: the boundary operation maps \(A\) to

  • \(B = \partial A = \partial \mathbb{H}^{n} = \{ (x_{1},\ldots,x_{n}) \in \mathbb{R}^{n} : x_{n} = 0 \}\) .

where

  • \(A\) is the domain set before \(\partial\) acts.
  • \(B = \partial A\) is the edge set after \(\partial\) acts.
  • \(n\) is a natural number with \(n > 0\).

example 4 [d] (Edge of an edge is empty — Emam; Nakahara) Apply \(\partial\) twice to a solid ball. Before the first application,

  • \(A = D^{3}\)

is the solid ball. After one application,

  • \(B = \partial A = S^{2}\)

is the sphere. After a second application the edge of that edge is empty:

  • \(\partial B = \partial^{2} A = \emptyset\) .

Emam: the boundary of a boundary is always vanishing; a closed manifold such as a circle has empty boundary, written \(\partial M = 0\), and \(\partial^{2} = 0\). The same holds for a sphere. Nakahara: the boundary of the solid ball \(D^{3}\) is the sphere \(S^{2}\), and the boundary of the sphere is an empty set.

where

  • \(A\) is the domain set before \(\partial\) acts.
  • \(B = \partial A\) is the edge set after one application of \(\partial\).
  • \(\partial^{2} A = \emptyset\) is the result after \(\partial\) acts on \(B\).

example 5 [d] (Before and after in Stokes’s theorem — Gowers) Before: a definite integral is taken over a domain set \(A = D\). After the boundary operation produces \(B = \partial D\), Stokes’s theorem moves that definite integral onto the edge set:

  • \(\displaystyle \int_{A} d\omega = \int_{B} \omega = \int_{\partial D} \omega\) .

The sets alone change from \(A\) to \(B = \partial A\); the equality says the two definite integrals have the same value.

where

  • \(A = D\) is the domain set before the move.
  • \(B = \partial D\) is the edge set after \(\partial\) acts.
  • both integrals are definite integrals.

157.2 Elementary Example

157.2.1 Simple

The boundary operation \(\partial\) maps a domain set \(A\) to its edge set \(B = \partial A\).

\[ A = \{ a,\ b,\ c,\ d \} \]

\[ B = \partial A = \{ a,\ b \} \]

where

  • \(A\) is the domain set before \(\partial\) acts.
  • \(B = \partial A\) is the edge set after \(\partial\) acts.
  • \(\partial\) is the boundary operation.

157.2.2 General

Nilpotence says the edge of an edge is empty: \(\partial^{2} = 0\). A solid ball maps to a sphere, then to the empty set.

\[ A = D^{3} \]

\[ B = \partial A = S^{2} \]

\[ \partial B = \partial^{2} A = \emptyset \]

where

  • \(D^{3}\) is the solid ball.
  • \(S^{2}\) is the sphere, the edge of \(D^{3}\).
  • \(\emptyset\) is the empty set.

157.3 References

  1. Emam, M. H. Covariant Physics. Oxford University Press, 2021. — \(\partial\) as a nilpotent operator; \(\partial^{2}=0\); line segment to endpoints.
  2. Gowers, T., Barrow-Green, J., & Leader, I. (eds.). The Princeton Companion to Mathematics. Princeton University Press, 2008. — boundary operation in Stokes’s theorem; adjoint of differentiation.
  3. Lee, J. M. Introduction to Topological Manifolds. Springer, 2011. — closed unit disk to circle; \(\mathbb{H}^{n}\) to \(\partial \mathbb{H}^{n}\).
  4. Nash, C., & Sen, S. Topology and Geometry for Physicists. Academic Press, 1983. — \(b(U) = \{ a, b \}\) for an interval.
  5. Nakahara, M. Geometry, Topology and Physics. Institute of Physics Publishing, 2003. — \(\partial D^{3} = S^{2}\) and \(\partial S^{2}\) empty.